Automated voice and speech labeling

Abstract

A system and method for voice and speech analysis which correlates a speaker signal source and a normalized signal comprising measurements of input acoustic data to a database of language, dialect, accent, and / or speaker attributes in order to create a transcription of the input acoustic data.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application claims priority to U.S. Provisional Application No. 61 / 617,884, filed on Mar. 30, 2012, the entirety of which is hereby incorporated by reference.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]The present invention relates to a speech data analysis system, and, more specifically, to a system that correlates speaker signal source and a normalized signal comprising measurements of input acoustic data to a database of language, dialect, accent, and / or speaker attributes in order to create a detailed transcription of the input acoustic data.[0004]2. Description of the Related Art[0005]Speech transcription is an evolving area of technology served by several disparate technologies targeted at subsets of the issue. Individual systems and applications focus on and attempt to solve their own problems, including speech-to-text, phrase and word recognition, language recognition, and speaker identification. However, ...

AI Technical Summary

Benefits of technology

The present invention provides a speech-to-text transcription system that does not require excessive training. The system uses speech decomposition and normalization methods to provide a normalized signal. It compares obtained data to a database of language, dialect, accent, and speaker attributes in order to produce a transcription of the speech. Other objects and advantages of the invention are to improve the accuracy and efficiency of speech-to-text transcription systems.

Problems solved by technology

However, each of these techniques applies only rudimentary signal processing techniques, and none are able to achieve high levels of accuracy without a large amount of training.

Although effective, domain training results in several limitations on the application of the approach, both in the specific speech domain and in how much confidence the user has in the product.

Additionally, in situations in which a significant amount of training is required, the time and effort required can be a substantial barrier to adoption by new users.

In addition to training requirements, ASR systems continue to suffer from less-than-perfect accuracy, with some estimating a current peak effectiveness of only 80-90%.

Although ASR systems can greatly increase productivity, the need to correct converted speech detracts from the possible productivity maximum.

Images